Would you like to see full default interoperability between #ATproto and #ActivityPub without a bridge?
General Discussion
47
Posts
8
Posters
2.5k
Views
-
There are a lot of application domains where multiple competing protocols prevent growth of the industry as a whole. I'd pick Internet of Things and instant messaging as two good examples.
-
There are others where a single protocol has become the overwhelming default, and agreeing on that protocol lets us move up the stack and produce incredible value for everyone. HTTP and SMTP are two good examples.
-
-
-
@julian @wjmaggos I also think that the difference between open protocols and proprietary protocols matters. Open protocols level the playing field, while commercial protocols concentrate power in one entity. Other players in the space have an incentive to focus on open alternatives to commercial protocols, and sometimes this can begin convergence. Having active or dominant commercial protocols in a space is a source of instability.
-
@julian @wjmaggos I also think that the difference between open protocols and proprietary protocols matters. Open protocols level the playing field, while commercial protocols concentrate power in one entity. Other players in the space have an incentive to focus on open alternatives to commercial protocols, and sometimes this can begin convergence. Having active or dominant commercial protocols in a space is a source of instability.
-
-
@evan @wjmaggos it's not my message. My message to technologists was "the key piece of interop for humans is identity" and I was told "webfinger isn't part of activitypub" and "domain names are how we do identity on the internet" (despite the idea that paying $25/year to get to manage DNS isn't an end-user-achievable thing in the world we live in) and after years of this I gave up. It makes me sad that Bluesky went with domain names because consistent format really matters, but 🤷♂️
@blaine @evan @wjmaggos on a tangent to this, it took me 10 minutes each to get a working instances of OSSN & Piwogo up on my website hosted on Bluehost servers, using Softaculous scripts, simply because they can be installed in directories on existing domains. Meanwhile, I have failed to date in self hosting Fediverse instances, limited by needing to register yet another domain name (which I've done anyway), and not figuring out secure port forwarding for a home Yunohost server.
-
@julian @wjmaggos I don't know if there is a good book about how SMTP won, by the way. It would be a great read. In hindsight, it seems inevitable, but in the late 1980s and early 1990s it was far from a sure thing. There were dozens of commercial network email protocols already in use, and a few interoperation standards, like X.400.
-
-
@julian @wjmaggos I don't know if there is a good book about how SMTP won, by the way. It would be a great read. In hindsight, it seems inevitable, but in the late 1980s and early 1990s it was far from a sure thing. There were dozens of commercial network email protocols already in use, and a few interoperation standards, like X.400.
-
@blaine @evan @wjmaggos on a tangent to this, it took me 10 minutes each to get a working instances of OSSN & Piwogo up on my website hosted on Bluehost servers, using Softaculous scripts, simply because they can be installed in directories on existing domains. Meanwhile, I have failed to date in self hosting Fediverse instances, limited by needing to register yet another domain name (which I've done anyway), and not figuring out secure port forwarding for a home Yunohost server.
@TotalSonic @blaine @wjmaggos you're right; that is a tangent.
-
@chris @blaine @wjmaggos it's funny that you say "without bridges" and "however that may happen" in the same breath. If we don't care how it happens, why rule out one particularly powerful technique for interoperability?
BridgyFed works GREAT. It's fucking amazing. Is there room for improvement? Absolutely. But it's a really good service.
-
@chris @blaine @wjmaggos it's funny that you say "without bridges" and "however that may happen" in the same breath. If we don't care how it happens, why rule out one particularly powerful technique for interoperability?
BridgyFed works GREAT. It's fucking amazing. Is there room for improvement? Absolutely. But it's a really good service.
-
@wjmaggos @chris @blaine one of my big problems with it too. I think @anewsocial are aware of that limitation and are working on some solutions.
-
@evan @julian @wjmaggos what would you consider to be the "application domain" of HTTP, and what needs to be built above it? ironically, i think if you look at the OSI model, then HTTP is already at the uppermost "layer 7", the application layer -- and attempts to build networks on top of HTTP are in effect virtualizing lower layers! so instead of a virtual 3-7 on top of HTTP's 7, why not use HTTP's 7 directly? how many nested virtualized app layers do we need?
-
@evan @julian @wjmaggos what would you consider to be the "application domain" of HTTP, and what needs to be built above it? ironically, i think if you look at the OSI model, then HTTP is already at the uppermost "layer 7", the application layer -- and attempts to build networks on top of HTTP are in effect virtualizing lower layers! so instead of a virtual 3-7 on top of HTTP's 7, why not use HTTP's 7 directly? how many nested virtualized app layers do we need?
-
-
@trwnh @julian @wjmaggos I am taking Computer Networks right now, and I asked the very same question. If we have application-layer protocols, what do we call the things like SOAP or ActivityPub that work on top of those protocols. The answer I got was that those are also layer 7. Once you get to layer 7, it's just 7 all the way up. Inverted turtle structure!
-
@wjmaggos @chris @blaine one of my big problems with it too. I think @anewsocial are aware of that limitation and are working on some solutions.
@evan @wjmaggos @chris @blaine @anewsocial
Appreciate all the kind words in the thread! That said, I still strongly believe opt-in is a feature, not a bug. The right solution isn't to force people onto networks they don't want to be a part of; it's to clean up the rough edges of unbridged interactions and content so it makes less of a difference.
Gli ultimi otto messaggi ricevuti dalla Federazione
-
Right, which ActivityPub.Space follows and auto categorizes, which ends up as a way for microblog content to make its way into the threadiverse. @strypey@mastodon.nzoss.nz
@silverpill@mitra.social sorry for interjecting 😁
-
-
> @strypey@mastodon.nzoss.nz said:
>
> We could instead try to agree on a separate place for all discussions about fediverse coordination (would it rub anyone up the wrong way if I described this as fediverse governance?). But it does makes sense to discuss a federated network in a federated network, rather than a separate, centralised forum.I keep harping on this one point because it just makes so much sense that it'd be silly not to. The new blood (new software devs implementing AP) always appear on fedi, not some third-party site, first.
To your main point, fedi on Mastodon is all just everybody talking in the same room. Certainly that's why it seems like there's so much meta-discussion about AP.
This isn't the case on the threadiverse, where discussions are segregated/categorized by community.
-
@mariusor awesome! ♥️
-
@pfefferle I added a PKCE task to my todo list:
-
@pfefferle if I'd come up with a scheme for them they would be related to pairs of activityType:objectType entries (instead of a global write), or maybe just the activityType
Ie, Create:Note, Create:Image, Undo, Delete.
-
@mariusor yea, I think the scopes and the PKCE part was the issue!
but finally... 😊
-
@pfefferle love it! 💗
GoActivityPub doesn't really do "scopes" because I couldn't think of a way store the permissions on the server side in a way that doesn't require custom logic to interact with vanilla ActivityPub objects.
